AI Hallucinations in the Legal Sector: Professional Risks and Insurance Implications

Judgment Exposes AI Risks in Legal Practice

In the recent case Ayinde, R v The London Borough of Haringey [2025], Mr Justice Ritchie uncovered a serious issue: the claimant’s barrister submitted written arguments supported by entirely fictitious case citations. The solicitor’s response—that these fake cases were merely "cosmetic errors" and did not require correction—deeply concerned the judge.

The judge emphasized that ensuring the accuracy of facts and legal grounds is a fundamental duty of the legal team, including solicitors. Presenting five fabricated cases was deemed professional misconduct, highlighting the critical importance of verifying legal authorities before submission.

Consequences and Regulatory Actions

Although the judicial review succeeded on its merits, the claimant’s legal team faced a £7,000 reduction in costs due to their conduct. Furthermore, the judge ordered the defendant to forward the hearing transcript to both the Bar Standards Board and the Solicitors Regulation Authority (SRA) for further scrutiny.

This judgment underscores the risks posed by AI systems that can generate plausible yet false information, commonly known as “AI hallucinations.” It serves as a warning that legal professionals must maintain traditional verification methods to authenticate citations and authorities, even when using AI tools.

AI in Legal Services: Progress and Precautions

The use of AI in legal services is expanding, with the SRA recently approving Garfield.Law Ltd, the first AI-driven law firm. However, the SRA has echoed concerns about the risks of AI hallucinations, especially in generating relevant case law—a task identified as particularly high-risk for large language models.

These challenges are not limited to legal professionals. The insurance market, especially firms providing professional indemnity (PI) coverage, faces exposure from potential AI-related errors. The SRA reports significant AI adoption, with three-quarters of the largest solicitors’ firms and 72% of financial services firms currently using AI tools. Legal services are expected to be among the most affected sectors over the next decade.

Hallucinations in Practice: A Cautionary Tale

In a striking example, a Minnesota judge criticized an expert witness specializing in technology-based deception research for submitting evidence drafted with ChatGPT that referenced non-existent academic studies. This case (Kohls and Franson v Ellison - Case No 24-cv-03754) highlights that even experts are vulnerable to AI hallucinations, raising concerns for all professionals relying on AI-generated content.

Practical Steps for Insurers and Legal Professionals

Insurers must proactively address the evolving risks linked to AI. Some have started updating policies to cover losses caused by underperforming AI tools, signaling growing awareness of AI-related liabilities. However, how the PI market will handle claims based on AI errors remains uncertain.

Legal professionals and insurers should maintain open communication about the AI tools in use before policy renewals and throughout policy periods. This ongoing dialogue is crucial to managing risks effectively and ensuring appropriate coverage.

• Verify AI-generated citations with traditional legal research methods.
• Discuss AI tool usage and related risks with insurers regularly.
• Stay informed about regulatory expectations and updates on AI in legal practice.

For legal professionals interested in strengthening AI-related skills and understanding AI risks, exploring targeted courses and certifications can be valuable. Resources such as Complete AI Training’s legal-focused courses offer practical guidance on integrating AI responsibly.